SEED Guide

Evaluating: Tasks & Tools

Steps

Hypothesis Testing

Hypothesis testing involves questions about the Desirability - Feasibility - Viability tryptic.

1. To test market risk (desirability), ask:

➢     Do customers want this solution/product/service?

➢     Define sample size.

2. To test infrastructure risk (feasibility), ask:

➢     Can we build it?

➢     Does the tech exist?

3. To test financial risk (viability), ask:

➢     Does the business model exist?

➢     Can we generate enough revenue? Will it make money?

Setting the stage for testing,

https://designthinking.nusselder.org/testing/

After establishing the WHAT, WHY, and HOW, consider the responses to these questions:

–Who will test your prototype?

–What exactly are you testing for?

–Where will you be testing?

Note especially the questions you will want to explore with the stakeholders:

➢     What resonates with you?

➢     What surprises you?

➢     What do you think is missing?

Storyboarding

https://devsquad.com/blog/storyboards-design-thinking

http://creatingminds.org/tools/storyboarding.htm

1. The storyboard you create, with the following six (6) sketches, can test your hypothesis visually.

➢     customer

➢     insight

➢     problem definition

➢     value proposition

➢     how it works

➢     competitive context

2. Carefully re-enact your proposals and observe the interaction of each sketch.

Test Cards

(see template below)

Strategyzer’s Test Card

1. Test your hypothesis by completing the following prompts for (1) your hypothesis, (2) testing method(s) and (3) metrics, and (4) criteria:

➢     We believe that (1).

➢     To verify that, we will (2) and measure (3).

➢     We are right if (4).

2. For each test card, include a name for the test, the duration, and the deadline as well as the team member the test card has been assigned to.

Test Grid planning

(see template below)

Use one quadrant for each question:

➢     What was bad?

➢     What was good?

➢     Did you discover any new problems?

➢     Did you discover any new ideas?

A/B testing

https://hbr.org/2017/06/a-refresher-on-ab-testing

A/B testing is over a century old but in an online context, it is now easier to reach users to discover responses to questions like “What is most likely to make people click/buy our product/register with our site?”

For this simple form of a randomized controlled experiment, assign different versions of a product or site to two randomly determined sets of users.

The objective is to determine which version influences your success metric.

Avoid these common mistakes:

➢     Do not react early to data. Instead, conclude the test and only then, consider the data.

➢     Do not get distracted by too many metrics. Instead, focus on the most relevant metrics.

➢     Do not react to false positive results. Instead, do enough re-testing to be sure.